User Fees for Roads: The Stockholm Experience

Ottawa’s recent decision to increase its emissions reduction target to 80% from 2012 levels by 2050 will require major changes to one sector in particular: transportation, currently responsible for 40% of the City’s community greenhouse gas emissions.

In a previous blog, Trevor Haché discussed how introducing congestion charges could help alleviate the City’s dependence on private vehicles and incidentally reduce the burden of road repair and maintenance. In this and subsequent blogs, I will look at experience with congestion charges in other cities.

As the previous blog mentioned, congestion charges have been (or are about to be) tried in various forms in a wide range of cities and countries, including a number of Canadian and U.S. cities, as well as in Europe and Asia. In addition to reducing congestion in inner cities, a major secondary benefit of these charges is revenue generation, and in particular creating revenue that can be used to offset the wear-and-tear of vehicle transport on the city’s roads—a major ongoing cost for both the city and the province. Congestion charges are also a very cost-effective way to incentivise the “modal shift” to public transport that Ottawa will need once its LRT network is implemented. And finally, they offer one solution to the problem of reducing greenhouse gas emissions from the transportation sector.

But exactly how effective are congestion charges, and what experience from other jurisdictions could Ottawa use to ensure that the introduction of such charges will be successful? This blog tries to answer these questions by looking at a one city that has conducted a detailed post-implementation evaluation.

The best known example of congestion charges is Stockholm, which introduced road pricing on a provisional basis in 2006, and then on a more permanent basis in 2007. The provisional phase was intended to give citizens a chance to experience the effect of congestion charges--at a time when public resistance to the idea was considerable—before agreeing to institute them for the longer term. Following that phase, a referendum was held on the subject, which passed by a substantial majority. The charging system consists of a cordon around the inner city, with a time‐differentiated toll being charged in each direction. According to a recent analysis, traffic across the cordon has been reduced by around 20%, leading to substantial congestion reductions in and around the city.

It is useful to compare Stockholm’s situation with Ottawa’s, since the first objection to such comparative analyses is that cities are unique and such programmes are therefore not easily replicated.

First, Stockholm’s population is nearly identical with Ottawa’s: approx. 1 million. However around 2/3 of the city’s inhabitants live in the inner city, an area of some 35 km2. The population within the “toll zone” or cordon is around 330,000, of which 60,000 commute to workplaces outside the zone.

By comparison, 81% of Ottawa’s population lived in the “inner core” in 2001, but this has declined significantly in recent years as new outlying areas were incorporated into the city, and is probably more like 60% (the proportion “inside the Green Belt”) at present. On the other hand, Ottawa’s employment is heavily concentrated in the inner city (if we define this as the area within a 35 km2 “ring” around the Parliament buildings), as therefore is commuting. The main target of a congestion charge in Ottawa would therefore be suburban residents working in the inner city, rather than inner city residents working outside the core.

The main difference between Stockholm and Ottawa is the availability of public transit. Ottawa Transpo’s bus service is relatively efficient and heavily utilised for inner-city commuting, but less so for suburban transit. The advent of LRT services over the next 10-15 years will change this picture dramatically, but even then the city will have far less transit availability than Stockholm currently has. According to the Centre for Transport Studies:

In Stockholm, “60‐65% of all motorized person trips to and from the city centre are made by transit. During rush hours, the share increases to 80%. The public transport system in the county of Stockholm consists of a subway network with 100 stations and over a million trips per day, a commuter rail network with 51 stations and nearly a quarter of a million trips per day, five light rail lines with 98 stations with a bit more than 100 000 trips per day, and an extensive bus network with nearly a million trips per day.”

Despite the availability of transit options, Stockholm’s rush hour traffic density is three times that of its off-hour density, so private vehicle congestion was definitely a serious problem.

The differences between Stockholm and Ottawa are therefore smaller than might be expected, with transit availability the main distinguishing feature. The notion of an inner-city “cordon” would also differ, as in Ottawa’s case the main peripheral roads—176, 417, 17—do not “ring” the major employment areas but rather run through them. Still, these roads do funnel the majority of commuter traffic in and out the city, and would therefore present a logical target for cordon-based pricing, subject to a more detailed study of traffic flows during commuter periods.

The lessons learned (and ideally transferable) from the Stockholm experience are that (i) even a small decrease in congestion—20% in their case—will be perceived positively by road users; and (ii) a gradual or provisional introduction of congestion charges is preferable to a sudden imposition. As well, the lack of strong transit alternatives at present may lead to greater initial user resistance in Ottawa than was the case in Stockholm. At the same time, there is clearly a strong need for incentivising the transition to public transit when it does come, and congestion charges are an effective way to do that.

It should be noted as well that the Stockholm experience had strong national government support, and the referendum was even timed to coincide with national elections. Subsequent to the elections, the city was able to use the referendum success as a way to leverage major transport funding from the federal government, arguing that by improving revenue from commuters, the city was able to pay for a portion of new road development itself. Although it may seem ironic and counter-intuitive that a cordon charge was used to pay for new roads, this did make sense within the very different context of Swedish coalition politics. And of course Stockholm, unlike Ottawa, did not need to use the additional revenue to repair old roads, which were in very good condition!

A few other benefits of the Stockholm charge:

1. Travel times for commuters decreased by as much as 2/3 for peak travel periods, although this varied quite a bit for other travel periods.

2. Transit usage increased by 4-5%. Note that there were some additions to the transit network, intended to provide an additional stimulus to encourage modal switching.

3. CO2 emissions for the wider Stockholm region decreased by 2-3%.

4. Attitudes towards congestion charges improved dramatically both during the trial period and after, suggesting that “familiarity breeds acceptability.”

A detailed examination of the history and impacts of the Stockholm congestion fee experience, written in 2014, can be found at: 

A cost-benefit analysis can found at: ;


Geoff Stiles

Carbon Impact Consultants

Share this page

Take action

Make a Donation
Sign up to Volunteer
Become a Member
Take Action!

Connect with us

Make a donation